Atypical Event-Related Potentials Revealed During the Passive Parts of a Go-Nogo Task in Autism Spectrum Disorder: a Case-Control Study Anne L

Høyland et al. Molecular Autism (2019) 10:10 https://doi.org/10.1186/s13229-019-0259-3

RESEARCH Open Access Atypical event-related potentials revealed during the passive parts of a Go-NoGo task in autism spectrum disorder: a case-control study Anne L. Høyland1,2* , Terje Nærland3,4, Morten Engstrøm5,6, Tonje Torske7, Stian Lydersen1 and Ole A. Andreassen4,8

Abstract Background: The core features of autism spectrum disorder (ASD) are easily recognizable in non-structured clinical and real-life situations. The features are often difficult to capture in structured laboratory settings, and the results from tests do not necessarily reflect symptom severity. We investigated neurophysiological processing in the passive parts of a cued Go-NoGo task, using the active parts of the test as a comparator. Methods: Forty-nine adolescents diagnosed with ASD and 49 typically developing (TD) adolescents (age 12–21 years) were included. Daily life executive function was assessed with the Behavior Rating Inventory of Executive Function (BRIEF). We applied a visual cued Go-NoGo task and recorded event-related potentials (ERPs). We investigated occipital N1, a component related to early perception of visual stimuli, and P3a, a fronto-central component related to switching of attention, in the passive and active parts of the test. Results: During the passive parts, the ASD group had statistically significantly longer N1 latency (p < 0.001, Cohens d= 0.75) and enhanced amplitude of P3a (p = 0.002, Cohens d= 0.64) compared to the TD, while no significant differences were observed in the active parts. Both components correlated significantly with the Behavioral Regulation Index of the BRIEF (partial correlation r = 0.35, p = 0.003). Conclusion: Delayed N1 response, indicating altered visual perception, and enhanced P3a response, indicating increased neural activation related to attention allocation, were found during the passive parts of a Go-NoGo task in ASD participants. These abnormal ERP signals in the non-structured settings were associated with everyday executive function, suggesting that neurophysiolocal measures related to atypical control of alertness and “hyper-awareness” underlie daily life dysfunction in ASD. Assessments during passive settings have a potential to reveal core neurobiological substrates of ASD. Keywords: ASD, ERP, Passive condition, N1, P3a

* Correspondence: [email protected] 1Department of Mental Health, Faculty of Medicine and Health Sciences, Regional Centre for Child and Youth Mental Health and Child Welfare, Norwegian University of Science and Technology, Klostergata 46, N-7030 Trondheim, Norway 2Department of Pediatrics, St. Olavs hospital, Trondheim University Hospital, Trondheim, Norway Full list of author information is available at the end of the article

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Høyland et al. Molecular Autism (2019) 10:10 Page 2 of 11

Background to neurobiological processes involved in cognition [33]. Autism spectrum disorder (ASD) is a neurodevelopmental There are numerous ERP studies in ASD, but there is a disorder with impaired reciprocal interaction and a large degree of variation in the findings. The occipital restricted pattern of behavior [1, 2]. ASD is now widely N1 wave, an early response to visual stimuli, is modu- accepted as a developmental neurobiological disorder with lated by selective attention [34]. The N1 amplitude is multifaceted etiology comprising genetic, environmental, enhanced in discrimination tasks [35], and studies have and gene-by-environment influences resulting in pertur- found alterations in the visual evoked potentials in bations in early brain development [3, 4]. Despite clear ASD [36]. The P3 has an early, fronto-central compo- central nervous system involvement, ASD is still purely nent, P3a, and a later and more posterior component behaviorally defined. The core behavioral features of ASD P3b [37, 38] and is elicited whenever a task conveyed are highly context-dependent, and assessment of the contextual information about an impending change [39, behaviors that define ASD has proven difficult to do in a 40]. In a meta-analysis of the P3 components, Cui et al. stringent and controlled manner. Typical behavior is [41] reported a tendency of reduced P3b amplitude among usually more apparent in complex real-life situations than ASD subjects. A striking feature in their analysis is the in standardized laboratory settings [5], suggesting that heterogeneity among the 32 studies included. Differences un-structured or passive test situations are needed to re- in amplitudes and latencies are attributed to variances in veal neurophysiological substrate for ASD. paradigm and the heterogeneity among ASD participants. Constant adaption to potentially relevant environmental Another component related to change detection is the events balanced against attention modulated by expectation mismatch negativity (MMN), typically elicited by auditory and desire or relevance is necessary to achieve goal-directed stimuli. Analogous response occurs also in other sensory behaviors [6, 7]. The ability to control sensory modalities [42]. Findings from studies of MMN in ASD responsiveness through gating mechanisms, i.e., filtering ir- have been inconsistent. As in studies on P3, both ampli- relevant or interfering stimuli or impulses, is fundamental tudes and latencies are found to be significantly larger for normal functioning [8]. Such information processing [43], in the normal range [44] and reduced [31, 45]. How- mediating selectivity is defined as attention [9–13]. Atypical ever, few studies have focused on ERPs during unstruc- attention modulation is suggested as a basic factor in the tured sessions of the cognitive experimental procedures. development of core features in ASD [9, 14]. Atypical sen- Difficulties with various aspects of executive function sory processing is repeatedly described [15, 16]andmay (EF) in everyday life among ASD subjects are consist- reflect core neuropathology underlying clinical symptoms ently observed by clinicians and family members [46– of ASD. Sensory processing dysfunction is common [17– 48]. The disparity between neuropsychological testing 20] and found in as much as 90% of children with ASD of EF and real-life EF is well known [5, 46, 49, 50]. [21]. Obligatory processing of task-irrelevant stimuli seems Neuropsychological examination is usually conducted to be a hallmark of autistic cognitive style, probably as a re- in a controlled, quiet, and structured setting, while the sult of insufficient top-down filtering [12, 22]. Several stud- real world is complex, noisy, unstructured, and at times ies support that individuals with ASD manifest unusual even chaotic [50]. Thus, a questionnaire for EF assess- neural responses to task-irrelevant features [6, 22–25], sug- ment in real-life setting, the Behavior Rating Inventory gesting that passive or unstructured test settings can reveal of EF (BRIEF), was developed [51]. BRIEF is suggested relevant neurobiological alterations. to capture distinct patterns of EF problems in individ- Studies on attention in ASD compared to typical de- uals with ASD [52, 53]. The cued Go-NoGo task is an veloping (TD) yield mixed results both regarding experimental task that enables the study of different performance and neurophysiological responses [26–29]. executive control processes as attention, reaction time, Both reduced and enhanced neural activation is found and inhibition of pre-potent responses [54]. The active in novelty-processing [25, 30–32]. This is suggested part of the test is when the participant either has to dependent of experiment context and seems different respond or inhibit a response to a stimulus. The passive for active (requiring a response) or passive (no re- parts are present after stimuli, when the participant nei- sponse) novelty detection. Keehn et al. found increased ther needs to respond nor inhibit a response. Although neural activation in ASD when given an active task but the passive parts may be differently related to the cog- reduced processing during passive tasks [9]. Deviant nitive task, we propose that it could reveal important processing during passive tasks could be related to electrophysiological abnormalities of ASD since core atypical control of alertness, over-focusing, and mal- autistic characteristics are particularly revealed during adaptive attention allocation. unstructured settings. Event-related potentials (ERPs) are small voltage os- We have previously reported no difference in behavior cillations measured at the scalp that are time-locked to performance among ASD and controls in a Go-NoGo task the processing of external events and can give insight [28] and similar ERPs during the active parts of the test Høyland et al. Molecular Autism (2019) 10:10 Page 3 of 11

[29]. However, the parents rating on the BRIEF indicate valid for the ASD diagnosis [57, 58]. ADI-R has been significant EF difficulties in the ASD group. Thus, our test found to discriminate well between ASD and non-ASD results in a structured experiment of executive control [59]. The ASD group in our study had markedly increased were not consistent with the reported real-life EF prob- scores on SCQ compared with TD (p <0.001). lems [50]. However, the passive parts of the Go-NoGo Eighteen (37%) individuals in the ASD group had neu- task may represent a setting that better mimicks the un- ropsychiatric comorbidity, all but one with attention structured real-life situations where abnormalities are well problems (Attention Deficit Disorder with or without documented. This is supported by several studies report- hyperactivity (ADHD)). Eight (16%) had more than one ing unusual neural responses to task-irrelevant features in comorbid diagnosis. Six (12%) had a diagnosis of epilepsy, ASD [6, 22–25]. The novel approach of the current study all but one with co-occurring ADHD. Twelve (25%) of the was to investigate the ERPs from both the non-active ASD individuals used medication regularly, four were on “breaks” (passive parts) of the Go-NoGo task and in the stimulants, two used atomoxetine and the six participants same test from the standard “task” parts (active). Few with epilepsy were on antiepileptic medication. studies have investigated this aspect of ERPs, as most test Intelligence quotients (IQs) were registered in the paradigms used in ASD are based on cognitive challenges ASD group. The IQs were mainly obtained during clin- and processing effort. We hypothesize that important in- ical assessments prior to the current study. Most of the formation about the core neurobiological substrate of participants were tested using the Wechsler tests [60], ASD could be revealed in the passive parts of the Go- while one participant was assessed using the Leiter test NoGo task. The aim of the present study was to identify [61] because of specific language problems. Some parti- differences in the electrophysiological processing in cipants were tested after recruitment into the current adolescents with ASD compared to typical developing ad- study, applying the Wechsler Abbreviated Scales of olescents (TD) during the passive parts of the Go-NoGo Intelligence [62]. When the difference between verbal task and investigate if these ERP components were associ- and performance IQs was ≥ 30, we did not calculate ated with real-life EF measured with the BRIEF indexes. full-scale IQ (FIQ). To be included in the study, verbal We hypothesized that ASD group would show abnormal (VIQ) or performance IQ (PIQ) had to be within the ERPs related to early visual processing (N1) and attention normal variation (≥ 70). allocation (P3a) during the passive parts of the Go-NoGo EF in everyday life were measured using the BRIEF task, since this represents an unstructured setting that bet- [51]. The BRIEF contains eight clinical scales that are ter reflects the real-life situation. grouped in a Behavioral Regulation Index (BRI) and a Metacognition Index (MI). The BRI comprises the Materials and methods child’s ability to modulate both behavior and emotional Participants control and the ability to move flexible from one activity Fifty adolescents with a confirmed diagnosis of ASD with- to another. The MI is associated with the ability for out intellectual disability from outpatients attending St. active problem solving and to initiate, organize, and Olavs Hospital, Trondheim, Norway, were included in the monitor their own actions [51]. The Global Executive study during 2013–2016. The sample consisted of 13 girls Composite (GEC) is a summary score that incorporates and 37 boys, aged 12–21 years, average 15.6 years. all eight clinical scales. T-scores of ≥ 65 are considered Forty-nine typically developing adolescents, matched to represent clinically significant areas. The ASD group for age and gender, were recruited from adjacent showed significantly increased GEC (mean (SD), ASD schools through invitations/bulletins to all students/ 67.8 (10.2), TD 42.2 (6.4), p < 0.001, Cohens d = 2.9). parents. In the invitation letter and recruitment posts, One of the participants in the ASD group had more we invited healthy adolescents. The parents confirmed than 70% omissions/commissions in the performance in writing that their child did not suffer from any data of the test and was excluded. The others, 49 ASD chronic disease or psychiatric problems presently or individuals and 49 TD, were included in the study. previously. Eighteen girls and 31 boys from 12 to 20 For detailed demographic information of the sample, years were included, average 15.6 years. see Additional file 1: Table S1. The ASD patients were diagnosed according to the ICD-10 [1] F.84 criteria for pervasive developmental Experimental task, electrophysiological recording, and disorder based on developmental information and clinical analysis assessments. To identify characteristics associated with Experimental task ASD, the parents of all participants completed the lifetime We used a visual-cued Go-NoGo task [63]. The catego- version of the Social Communication Questionnaire ries of visual stimuli (see Additional file 1: Figure S1, (SCQ) [55]. This questionnaire is based on the Autism http://www.mitsar-medical.com/eeg-software/qeeg-soft- Diagnostic Interview-Revised (ADI-R) [56] and is found ware/download.html) included 15 pictures of each Høyland et al. Molecular Autism (2019) 10:10 Page 4 of 11

category: animals, plants, and humans. All participants Electrophysiological recordings completed 300 trials. Each trial consisted of a subse- Electroencephalogram (EEG) was recorded using a Mitsar quent pair of stimuli (S1 and S2). When S1 was a cue, (http://www.mitsar-medical.com) EEG system with a 19- the subsequent S2 might require a response and the re- channel tin electrode cap (Electro-cap International, Eaton, cording after the S2 was assigned the active part of the OH, USA). The electrodes were placed according to the test. The passive parts of the test were all situations not international 10-20-system. The input signals were refer- requiring a response, that means after all S1 and after S2 enced to earlobe electrodes and filtered between 0.5 Hz when S1 was a non-cue (plant). In the non-cue trials, 50% and50Hzanddigitizedatasamplingrateof500Hz.Im- of the S2 were accompanied by a novel sound. The com- pedance was kept below 5 kΩ for all electrodes. Quantita- bination of auditory and visual stimuli complicated our in- tive data were obtained from the WinEEG software (http:// terpretation of ERPs and we excluded the results from this www.mitsar-medical.com) in common average montage part of the test. S1 and S2 were presented for 100 ms with prior to data processing. Eye blink artifacts were cor- an 1100 ms inter-stimulus interval and an inter-trial inter- rected by zeroing the activation curves of individual in- val of 3000 ms. The trials were grouped into blocks sepa- dependent components corresponding to eye blinks. In rated by short breaks. In each block, a unique set of five addition, epochs of the filtered EEG with excessive pictures from each picture category was selected. Each amplitude (> 100 μV) and/or slow (> 50 μVinthe0–1 block consisted of a pseudo-random presentation of 100 Hz-band) and excessive fast (> 35 μVinthe20–35 stimulus pairs with equal probability for each trial cat- Hz-band) frequency activity were automatically ex- egory. The participants were instructed to respond by cluded from further analysis. All participants also had a pressing a button with their right index finger as quickly six-minute resting EEG registration and a specialist in as possible without making mistakes in all Go trials and clinical neurophysiology examined the recordings and otherwise refrain from responding. They were also found no epileptic activity. instructed that a sound would turn up that they should The ERPs for each individual were based on averaging ignore. the trials of the respective task condition with correct During the task, participants were seated in a comfort- response after artifact correction. The number of able chair 1.2 m from the computer screen. The pictures artifact-free trials averaged were (mean (± SD, range)) (size approximately 20 × 15 cm) were presented in the TD 269 (± 22.4, 191–300), ASD 261 (± 37.9, 109–295). middle of an 18-in monitor using the Psytask (http:// This makes a non-significant difference in averaged tri- www.mitsar-medical.com/eeg-software/qeeg-software/ als. The ERPs were measured by convention; N1 as the download.html)) software (from Bio-medical, Clinton averaged peak amplitude and latency through O1 and Township, Michigan USA). The time interval from the O2 and P3a as mean amplitude in Cz, both in time win- presentation of the second stimulus to the response, re- dow chosen from the grand average file for all partici- action time (RT), and intra-individual reaction time vari- pants (Table 2). The topography of the P3a component ability (RTV) was registered by the software. The ERPs is illustrated in Fig. 1. The paradigm in our study does were averaged through trials with correct responses. The not involve a sequence of repetitive stimuli and does software also registered omissions and commissions. not elicit a MMN-component. Task performance and ERPs are previously reported (Table 1), for more details, see Høyland et al. [28]. Study design and outcomes The primary outcome for the current study was the ERPs N1 and P3a elicited during passive conditions in a Table 1 Performance in task and ERPs in a visual-cued Go- visual cued Go-NoGo task. We also investigated if these NoGo task [29] ERPs were associated to executive function as measured TD ASD by the BRIEF. Mean (SD) Mean (SD) p value Cohen’s d Statistical analysis n 49 49 We compared the N1 amplitude, N1 latency, and mean Reaction time 330.5 (62.0) 338.7 (65.2) 0.53 0.13 P3a amplitude between ASD and TD using Student’s t RTV 10.0 (3.7) 9.9 (3.6) 0.86 0.04 Omissions 3.5 (3.9) 3.7 (4.9) 0.75 0.06 Table 2 Event-related potentials (ERPs)—recordings Comissions 1.5 (1.9) 1.5 (1.9) 0.96 0 Time window Electrodes – Cue P3 (Pz) 4.41 (2.26) 4.41 (2.83) 0.99 0 N1 peak amplitude and latency 140 210 ms O1 and O2 – P3 NoGo (Cz) 11.66 (4.18) 11.94 (5.99) 0.92 0.05 P3a mean amplitude 250 320 ms Cz Cz P3 Go (Pz) 9.24 (4.15) 9.36 (3.00) 0.36 0.03 central midline electrode O1/2 respectively left and right occipital electrode Høyland et al. Molecular Autism (2019) 10:10 Page 5 of 11

Fig. 1 Brain mapping of P3a amplitude visualized as difference between ASD and TD, 174 ms after non-cue stimulus 1 test. In addition, we adjusted these analyses for age using experiment and sample, we adjusted the p values using linear regression, giving practically identical results (data the Benjamini-Hochberg procedure [64] to preserve the not shown). We calculated a receiver operating character- false discovery rate (FDR). istic (ROC) curve to estimate the sensitivity and specificity of the analyzed ERPs. We calculated the correlation Results between the ERP components and age using the Pearsons During the passive conditions, the N1 latency and the correlation coefficient. We also calculated the partial cor- mean P3a amplitude were significantly increased in ASD relation between these ERP components and the BRIEF (Table 3, Figs. 2 and 3). This was independent of stimu- indexes adjusted for age. lus relevance and in all situations where no action was Normality of residuals was checked by visual inspec- required (passive parts), i.e., after both cue and non-cue tion of Q-Q plots. Statistical analyses were carried out in S1 and passive S2, pointing in the direction of a more IBM SPSS Statistics 25.0. Since these analyses are based basic perceptual disturbance. By visual inspection of the on incidental discovery in the first analysis of the current grand average file, the P1 component seemed similar in

Table 3 N1 and P3a in passive parts of the test TD ASD Mean (SD) Mean (SD) n =49 n =49 p value Cohen’s d Passive phase after non-cue S1 N1 amplitude1 2.30 (4.86) 1.77 (5.13) 0.66 0.11 (S1 = plant) N1 latency1 168.3 (15.7) 181.3 (18.9) 0.003* 0.75 Mean P3a Cz − 0.99 (1.91) 0.25 (1.99) 0.005* 0.64 Mean P3a Fz − 2.57 (1.71) − 1.93 (1.92) 0.14 0.35 Passive phase after N1 amplitude1 2.37 (5.42) 1.72 (5.10) 0.66 0.12 S2 N1 latency1 165.4 (16.5) 179.3 (19.5) 0.003* 0.77 (S1 and S2 = plant) Mean P3a Cz − 0.86 (1.87) 0.35 (2.46) 0.014* 0.55 Mean P3a Fz − 3.00 (2.1) − 1.54 (2.0) 0.003* 0.82 Passive phase after cue S1 N1 amplitude 2.39 (4.95) 1.95 (4.91) 0.66 0.09 (S1 = animal) N1 latency 163.9 (15.4) 179.6 (20.5) 0.003* 0.87 Mean P3a Cz − 1.44 (1.85) − 0.36 (2.34) 0.024* 0.51 Mean P3a Fz − 1.29 (1.93) − 0.83 (1.85) 0.31 0.24 1Amplitude in microvolts and latency in milliseconds *Benjamini-Hochberg adjusted p values, significant at 0.05 level after adjustment for 12 hypotheses The data of N1 from O1 and O2 are averaged Høyland et al. Molecular Autism (2019) 10:10 Page 6 of 11

Fig. 2 Event-related potentials in left (O1) and right (O2) occipital electrodes after non-cue stimulus 1 in autism spectrum disorders (ASD) and typically developing (TD). Amplitude in microvolts (μV), latency in milliseconds (ms) the two groups (see Fig. 2). We also investigated the Both N1 latency and P3a were positively correlated with potential P3b after non-cue S1 and found no significant parent-rated measures of everyday executive function difference (see Additional file 1: Table S4). The N1 was (BRIEF) adjusted for age. When splitting into the indexes dependent of age, with negative correlations in both BRI and MI, we found a highly significant correlation TD and ASD (TD r = − 0.41, p = 0.004; ASD r= 0.52, between these ERP components and the BRI index, see p < 0.001), P3a showed no correlation with age (TD Table 4. r = 0.06, p = 0.66; ASD r= 0.03, p = 0.86). The N1 ampli- The relation between N1 latency/P3a amplitude tudes showed no significant differences between the ASD and BRI including ASD subtypes are illustrated in and TD groups in any of the recorded conditions (Table 3). Additional file 1: Figures S2 and S3. The ASD group The P3a has, as shown from the mapping (see Fig. 1), a included participants with comorbid ADHD. The distribution fronto-centrally, with main peak around Cz. case-control differences in ERPs and the correlations The ROC curve showed an area under curve (AUC) of to BRIEF were substantially the same after excluding 0.7 for both components, see Fig. 4. We chose cut points these participants, see Additional file 1:TablesS2 giving sensitivities as close to 80% as possible. We obtain and S3. a sensitivity of 80% and a specificity of 60% when setting We also investigated the same ERP components in a the cut-off point of N1 to 168.5 ms. Correspondingly, corresponding test with pictures of emotional faces setting the cut-off point of P3a to − 1.32 μVgivesa instead of animals/plants and obtained substantially the sensitivity of 80% and a specificity of 47%. same findings, see Additional file 1: Table S4.

Fig. 3 Event-related potentials in central midline electrodes (Fz and Cz) after non-cue Stimulus 1 in autism spectrum disorders (ASD) and typically developing (TD). Amplitude in microvolts (μV), time in milliseconds (ms) Høyland et al. Molecular Autism (2019) 10:10 Page 7 of 11

Fig. 4 ROC curves for N1 latency and P3a amplitude in passive phase after non-cue S1

Discussion many individuals with ASD show in non-structured set- The main finding of the present study was significant tings, which is a core feature of social human interac- case-control ERP differences in a cued Go-NoGo task tions. Thus, the current findings suggest that the during the passive parts of the test. The ASD group had identified atypical neurophysiological mechanisms may a delayed occipital N1-component and increased fronto- also underlie everyday dysfunction in ASD. Lastly, it is central P3a-amplitude in the passive parts of the task, possible that the findings during the passive part of the while results in the active part were without significant neuropsychological test may be a promising candidate differences in line with previous reports [28, 29]. These for biomarker of ASD which may have a potential as as- abnormal ERP signals were associated with everyday ex- sessment of change. ecutive function, suggesting that neurophysiolocal mea- Our findings are in line with the assertion that abnor- sures related to atypical control of alertness and “hyper- mal processing stimuli is a key feature of ASD cognitive awareness” underlie daily life dysfunction in ASD. Taken style [22]. This suggests that ASD individuals have par- together, the present findings suggest that assessments ticular problems with the ability to gate sensory informa- during passive experimental settings reveal core neuro- tion. Keehn et al. [9] discuss how aberrant attentional biological substrates of ASD. mechanisms can be linked to the emergence of core Executive dysfunction in everyday life is typical for ASD symptoms. Top-down modulation of attention in ASD subjects and contributes substantially to the degree active or passive tasks influence levels of alertness and, of disability [46–48]. However, in the laboratory, both thus, processing of new information [9]. The increased performance and electrophysiological measures of EF N1-latency might reflect increased attentional load and may be equal to TD adolescents [5, 27, 29]. Interestingly, processing effort of both important and unimportant in the passive part of the cued Go-NoGo task, the ASD stimuli, and be a neurophysiological correlate of “hyper-- group differed from the controls. The finding that sig- awareness,” impairing the ability to suppress irrelevant nificant case control differences are revealed when no or interfering stimuli as described by Garavan [8]. In real action is triggered (passive parts) is consistent with par- life, this can reflect difficulties in ignoring distracting in- ental and clinicians’ experience of the increased struggles formation. Friedman and Miyake [65] sought to examine the association between inhibition related functions and Table 4 Partial correlations of ERPs in passive phase after non- found that resistance to distractor interference was cue S1 with BRIEF included indexes, controlled for age closely related with other components of everyday EF, BRIEF total BRI MI such as task-switching ability and cognitive failures. An n 73 74 74 increased P3a amplitude mirrors more effort involved in N1 latency 0.29 0.35* 0.24 allocation or orientation to novelty and change. Restrict- ive and repetitive behavior is one of the two diagnostic P3a 0.21 0.35* 0.11 domains in ASD-criteria in DSM-5 [2] and one of the *Significant at 0.01-level BRIEF total score and indexes. Behavior Regulation Index (BRI) and most striking clinical features of the neurodevelopmental Metacognition Index (MI) disorder. The resistance to change and insistence of Høyland et al. Molecular Autism (2019) 10:10 Page 8 of 11

sameness, aspects of repetitive behavior, may very well comorbidity in ASD [71]. A range of psychiatric conditions be coping mechanisms to aberrant perception. are found to show altered ERPs [72, 73], and disentangling Visual stimuli evoke neural activity in the visual cor- the effects of comorbidity is necessary. In our previous tex that would be captured by occipital electrodes. study [29], we reported increased response preparation and Classification of the stimulus is obligatory to obtain ac- enhanced conflict monitoring in adolescents with ASD dur- ceptable performance in the Go-NoGo task, initiating ing a Go-NoGo task. We interpreted this as a possible link top-down attentional influence on this early visual to the feature “Insistence of sameness”,butitmayalsobe component. The N1-attention effect is shown to be the related to enhanced performance and superior academic same for both target and non-target stimuli, consistent skills [74]. This was also found only in adolescents older with a simple modulation of feedforward sensory activity than 16 years or participants without comorbid ADHD and [66]. Attention to stimuli, and not passive watching, en- may, therefore, be linked to subgroups of ASD. Inconsistent hances N1 amplitude [66] and is expected to affect N1 in with the results in our previous paper [28, 29], excluding our study. We did not find differences in N1-amplitude, the participants with ADHD or specific age groups did not reflecting similar response in both groups. Reduced sen- affect N1 latency and P3a amplitude. This may indicate that sory gating might contribute to sensory overload and the abnormalities related to sensory perception and novelty experienced hypersensitivity [67]. Increased N1-latency detection are more closely connected to core features of may be associated with greater complexity [68, 69], dem- ASD and less to concurrent developmental disorders. onstrating an association between latency and processing The sensitivity and specificity of classification results effort. Enhanced N1-latency may therefore reveal both ab- based on the components N1 and P3 suggests that the errant sensory processing and altered attentional effect. current findings should be replicated in independent We found age-related changes of N1-latency in the same samples. Thus, if the paradigm is further improved, range in both TD and ASD, indicating similar maturation these ERPs may be developed into biomarkers of ASD. processes. There is a clear need for direct and stringent assess- The fronto-central P3a-component is described to re- ments of ASD that can be used as a measure of change; flect orientation to information about an impending the ERPs elicited in this procedure may be a promising change in the task [39]. Previous findings on P3a in ASD candidate for such assessment. are inconsistent. Gomot et al. [32] described enhanced We found a positive correlation of ERPs related to per- P3a in the ASD group, while Jeste and Nelson found ception and attention orientation and BRIEF, most mark- reduced P3a [70]. Our task requires discrimination, and edly to the BRI. BRI comprises the ability to modulate thus, a P3a is expected in all participants. In contrast to both behavior and emotional control which very well Keehn et al. [9], P3a showed a correspondingly signifi- can be affected by the hyper-awareness related to aber- cantly enhanced amplitude in all passive conditions in rant perception and attention allocation. The correla- ASD in our study. The P3b elicited in the active parts tions between ERP components and BRIEF scores were (Go and NoGo) were similar between the two groups of weak (r = 0.35). As suggested by the Additional file 1: participants [29]. Also the MMN literature has divergent Figures S2 and S3, this is probably due to a high variance findings in the ASD group, both enhanced and normal among different subgroups. The presence of subgroups and reduced MMN is described [45]. Both P3a and may induce noise and confound the present results. MMN are linked to sensory processing [32]. The coex- However, it may also indicate interesting phenomena re- istence of atypical sensory processing and deviant atten- lated to the pathophysiology of ASD which should be tional salience detection and responses to change will investigated in larger studies with statistical power to cultivate need of predictability and, thus, insistence of reliably identify relevant subgroups. Other aspects of EF sameness and resistance to change [30, 31]. as inhibition of prepotent responses in ADHD will also Our findings are divergent from the recent meta-ana- contribute to BRI and affect the variability of the results. lysis of P300 in ASD, where Cui et al. [41]reportnodiffer- Cognitive aspects captured by the MI seem more spared ences in P3a amplitude. Variances in paradigm and also in ASD and are less correlated to these ERPs. whether recording in the active or passive condition may contribute to this divergence [9]. In Adam and Jarrolds Strengths and limitations of the study study of inhibition [24], they found that children with All participants were tested by the same technician in ASD had difficulties inhibiting irrelevant stimuli but not the same lab to reduce variations caused by testing pre-potent responses, indicating a greater tendency to conditions. All participants initially included except process interfering distractors. one completed the task to satisfaction and was kept in As described in the “Materials and methods” section, our the final sample. The results presented were based on ASD sample included 18 individuals with neuropsychiatric an unexpected finding in a secondary analyses and the comorbidity. This is in line with other studies of complexity of the task used is not optimal for this Høyland et al. Molecular Autism (2019) 10:10 Page 9 of 11

research question. However, the findings are strong Additional file and there is no reason to doubt the main outcome of the study. The results should be replicated in inde- Additional file 1: Table S1. Demographics; number (n) and mean ± SD. Table S2. ERPs in TD compared with ASD without comorbid ADHD pendent samples using a simpler and more targeted (ASD-ADHD). Table S3. Partial correlations of ERPs in participants without paradigm. As the studies in this research-field have ADHD with BRIEF included subscales, adjusted for age. Table S4. ERPs in provided large differences in mean and SD, with a other conditions and recordings from other electrodes. Figure S1. Task stimuli. Figure S2. Relation between N1 latency and the Behavioral large variation in methodology, we were not able to Regulation Index (BRI) including ASD subtypes. Figure S3. Relation perform any power-analysis. Prior to the study, we de- between P3a amplitude and the Behavioral Regulation Index (BRI) cided to include 50 ASD adolescents and 50 matched including ASD subtypes. (DOCX 390 kb) controls based on reported studies applying similar ERP experimental paradigms and methodology [54]. Abbreviations ADHD: Attention deficit disorder with or without hyperactivity; ADI-R: Autism We included patients previously diagnosed with ASD, but Diagnostic Interview-Revised; ASD: Autism spectrum disorder; BRI: Behavioral did not repeat the diagnostic assessment. However, we Regulation Index; BRIEF: Behavior Rating Inventory of Executive Functions; underpinned the diagnosis by parent information through EEG: Electroencephalogram; EF: Executive functions; ERP: Event-related potentials; FIQ: Full-scale intelligence quotient; GEC: Global Executive the SCQ. The distribution between the diagnostic sub- Composite; IQ: Intelligence Quotient; MI: Metacognition Index; PDD- groups shows an overrepresentation of Pervasive Develop- NOS: Pervasive Developmental Disorder-Not Otherwise Specified; mental Disorder-Not Otherwise Specified (PPD-NOS) in PIQ: Performance intelligence quotient; RT: Reaction time; RTV: Intra- individual reaction time variability; SCQ: Social Communication the participants under the age of 16 years, but there were Questionnaire; SD: Standard Deviation; TD: Typical Developing; VIQ: Verbal no significant differences in the ASD symptoms as assessed Intelligence Quotient by SCQ. We did not perform tests to estimate IQs for TD, but the parents of our control group reported no learning Acknowledgements The authors thank all the participants and their parents for providing the problems or psychiatric problems, and they were recruited necessary information. This study is part of the BUPgen study group and the from school children with normal school performance. In- research network NeuroDevelop (South East Norway Regional Health dividuals with classical autism typically have significantly Authority). The authors thank Professor Juri Kropotov, Institute of the Human Brain, Russian Academy of Sciences, Saint Petersburg, Russia, for technical lower verbal IQs compared to performance IQs, although assistance and neuropsychologist Geir Øgrim, Neuropsychiatric Unit, Østfold this varies within the ASD group. This situation also makes Hospital Trust, Fredrikstad, Norway for assistance in interpreting the ERP-data. it challenging to match a control group [75]. We used the Funding BRIEF, a parent-report measure, as a description of the This study is funded by the Liaison Committee between the Central Norway presence of executive dysfunction in the participants. Re- Regional Health Authority (RHA) and the Norwegian University of Science search indicates that disagreement exists between perform- and Technology (NTNU). The project was supported by the Research Council of Norway (Grant #213694, #223273) and the KG Jebsen Foundation. ance-based tests and parent-report measures of EF [76]. We used parent-report BRIEF for all participants even Availability of data and materials though some of them were over 18 years old. This is be- The datasets analyzed during the current study are available from the cause we had information that all participants still lived corresponding author on reasonable request. with their parents and we wanted to use the same BRIEF Authors’ contributions questionnaire across age groups. Performance-based mea- ALH, OAA, and TN have contributed to the design of the study, the analyses surements of EF could have contributed to a broader evalu- and the interpretation of the data. SL and TT have contributed to analyses and the interpretation of data. ALH and ME have carried out the EEG ation of executive dysfunction in the participants. analyses and contributed to the interpretation of the data. All authors drafted and revised the manuscript and gave their final approval to the Conclusion version published.

Our findings of aberrant ERP signals in ASD during the Ethics approval and consent to participate passive part of a cued Go-NoGo task suggest altered visual The study was approved by the Norwegian Regional Committee for Medical perception (delayed N1) and increased neural activation and Health Research Ethics South East (2013/1236/REK South-East). Written informed consent was obtained from participants and/ or parents necessary related to attention allocation (enhanced P3a). Both com- due to age. ponents correlate significantly to the Behavioral Regula- tion Index of the BRIEF, suggesting relevance for real-life Competing interests dysfunctions. The occipital N1 latency showed a fairly The authors declare that they have no competing interests. high sensitivity and specificity for the ASD diagnosis. ’ Together, our results suggest that abnormal attention allo- Publisher sNote “ ” Springer Nature remains neutral with regard to jurisdictional claims in cation, atypical control of alertness and hyper-awareness, published maps and institutional affiliations. are key pathophysiological features of ASD. Further, the results indicate that assessments during the passive parts Author details 1Department of Mental Health, Faculty of Medicine and Health Sciences, of testing is needed to reveal important information of Regional Centre for Child and Youth Mental Health and Child Welfare, core neuropathology of ASD. Norwegian University of Science and Technology, Klostergata 46, N-7030 Høyland et al. Molecular Autism (2019) 10:10 Page 10 of 11

Trondheim, Norway. 2Department of Pediatrics, St. Olavs hospital, Trondheim 22. Belmonte MK. Obligatory processing of task-irrelevant stimuli: a hallmark of University Hospital, Trondheim, Norway. 3NevSom, Department of Rare autistic cognitive style within and beyond the diagnosis. Biol Psychiatry Disorders and Disabilities, Oslo University Hospital, Oslo, Norway. 4NORMENT, Cogn Neurosci Neuroimaging. 2017;2(6):461–3. KG Jebsen Centre for Psychosis Research, University of Oslo, Oslo, Norway. 23. Gomot M, Wicker B. A challenging, unpredictable world for people with 5Department of Neurology and Clinical Neurophysiology, St. Olavs Hospital, autism spectrum disorder. Int J Psychophysiol. 2012;83(2):240–7. Trondheim University Hospital, Trondheim, Norway. 6Department of 24. Adams NC, Jarrold C. Inhibition in autism: children with autism have Neuromedicine and Movement Science, Norwegian University of Science difficulty inhibiting irrelevant distractors but not prepotent responses. J and Technology, Trondheim, Norway. 7Division of Mental Health and Autism Dev Disord. 2012;42(6):1052–63. Addiction, Vestre Viken Hospital Trust, Drammen, Norway. 8Division of Mental 25. Orekhova EV, Stroganova TA. Arousal and attention re-orienting in autism Health and Addiction, Oslo University Hospital, Oslo, Norway. spectrum disorders: evidence from auditory event-related potentials. Front Hum Neurosci. 2014;8:34. Received: 7 May 2018 Accepted: 8 February 2019 26. Allen G, Courchesne E. Attention function and dysfunction in autism. Front Biosci. 2001;6:D105–19. 27. Tye C, Asherson P, Ashwood KL, Azadi B, Bolton P, McLoughlin G. Attention and inhibition in children with ASD, ADHD and co-morbid ASD + ADHD: an References event-related potential study. Psychol Med. 2014;44(5):1101–16. 1. ICD-10 WHO. ICD-10: International statistical classification of diseases and 28. Hoyland AL, Naerland T, Engstrom M, Lydersen S, Andreassen OA. The health-related problems. Geneva: WHO; 1992. relation between face-emotion recognition and social function in 2. DSM-5 APA. Diagnostic and statistical manual of mental disorders (DSM-5®): adolescents with autism spectrum disorders: a case control study. PLoS American Psychiatric Pub; 2013. One. 2017;12(10):e0186124. 3. Minshew NJ, Sweeney J, Luna B. Autism as a selective disorder of complex 29. Hoyland AL, Ogrim G, Lydersen S, Hope S, Engstrom M, Torske T, Naerland information processing and underdevelopment of neocortical systems. Mol T, Andreassen OA. Event-related potentials in a cued Go-NoGo task Psychiatry. 2002;7(Suppl 2):S14–5. associated with executive functions in adolescents with autism spectrum 4. Bowman LC, Varcin KJ. The promise of electroencephalography for disorder; a case-control study. Front Neurosci. 2017;11:393. advancing diagnosis and treatment in neurodevelopmental disorders. Biol 30. Gomot M, Belmonte MK, Bullmore ET, Bernard FA, Baron-Cohen S. Brain Psychiatry Cogn Neurosci Neuroimaging. 2018;3(1):7–9. hyper-reactivity to auditory novel targets in children with high-functioning 5. Kenworthy L, Yerys BE, Anthony LG, Wallace GL. Understanding executive autism. Brain. 2008;131(9):2479–88. control in autism spectrum disorders in the lab and in the real world. 31. Gomot M, Bernard FA, Davis MH, Belmonte MK, Ashwin C, Bullmore ET, Neuropsychol Rev. 2008;18(4):320–38. Baron-Cohen S. Change detection in children with autism: an auditory 6. Keehn B, Nair A, Lincoln AJ, Townsend J, Muller RA. Under-reactive but event-related fMRI study. NeuroImage. 2006;29(2):475–84. easily distracted: an fMRI investigation of attentional capture in autism spectrum disorder. Dev Cogn Neuroscience. 2016;17:46–56. 32. Gomot M, Blanc R, Clery H, Roux S, Barthelemy C, Bruneau N. Candidate 7. Green SA, Hernandez L, Bookheimer SY,DaprettoM.Saliencenetwork electrophysiological endophenotypes of hyper-reactivity to change in – connectivity in autism is related to brain and behavioral markers of autism. J Autism Dev Disord. 2011;41(6):705 14. sensory overresponsivity. J Am Acad Child Adolesc Psychiatry. 2016; 33. Luck SJ. An introduction to the event-related potential technique: MIT press; 55(7):618–26 e611. 2014. 8. Garavan H, Ross T, Stein E. Right hemispheric dominance of inhibitory 34. Hopfinger JB, West VM. Interactions between endogenous and exogenous – control: an event-related functional MRI study. Proc Natl Acad Sci. 1999; attention on cortical visual processing. NeuroImage. 2006;31(2):774 89. 96(14):8301–6. 35. Vogel EK, Luck SJ. The visual N1 component as an index of a discrimination – 9. Keehn B, Müller R-A, Townsend J. Atypical attentional networks and the process. Psychophysiology. 2000;37(2):190 203. emergence of autism. Neurosci Biobehav Rev. 2013;37(2):164–83. 36. Kovarski K, et al. Brief report: early VEPs to pattern-reversal in adolescents – 10. Marco EJ, Hinkley LB, Hill SS, Nagarajan SS. Sensory processing in autism: a and adults with autism. J Autism Dev Disord. 2016;46(10):3377 86. review of neurophysiologic findings. Pediatr Res. 2011;69(5 Pt 2):48R–54R. 37. Nieuwenhuis S, De Geus EJ, Aston-Jones G. The anatomical and functional 11. Johnston WA, Dark VJ. Selective attention. Annu Rev Psychol. 1986;37(1):43–75. relationship between the P3 and autonomic components of the orienting – 12. Katsuki F, Constantinidis C. Bottom-up and top-down attention: different response. Psychophysiology. 2011;48(2):162 75. processes and overlapping neural systems. Neuroscientist. 2014;20(5):509–21. 38. Jeon YW, Polich J. P3a from a passive visual stimulus task. Clin Neurophysiol. 13. Macaluso E, Noppeney U, Talsma D, Vercillo T, Hartcher-O’Brien J, Adam R. 2001;112(12):2202–8. The curious incident of attention in multisensory integration: bottom-up vs. 39. Barcelo F, Escera C, Corral MJ, Periáñez JA. Task switching and novelty top-down. Multisensory Res. 2016;29(6-7):557–83. processing activate a common neural network for cognitive control. J Cogn 14. Keehn B, Lincoln AJ, Müller RA, Townsend J. Attentional networks in Neurosci. 2006;18(10):1734–48. children and adolescents with autism spectrum disorder. J Child Psychol 40. Barceló F, Periáñez JA, Knight RT. Think differently: a brain orienting Psychiatry. 2010;51(11):1251–9. response to task novelty. Neuroreport. 2002;13(15):1887–92. 15. Rogers SJ, Ozonoff S. Annotation: what do we know about sensory 41. Cui T, Wang PP, Liu S, Zhang X. P300 amplitude and latency in dysfunction in autism? A critical review of the empirical evidence. J Child autism spectrum disorder: a meta-analysis. Eur Child Adolesc Psychol Psychiatry. 2005;46(12):1255–68. Psychiatry. 2017;26(2):177–90. 16. Robertson CE, Baron-Cohen S. Sensory perception in autism. Nat Rev 42. Czigler I, Weisz J, Winkler I. ERPs and deviance detection: visual mismatch Neurosci. 2017;18(11):671. negativity to repeated visual stimuli. Neurosci Lett. 2006;401(1–2):178–82. 17. Behrmann M, Thomas C, Humphreys K. Seeing it differently: visual 43. Ferri R, Elia M, Agarwal N, Lanuzza B, Musumeci SA, Pennisi G. The mismatch processing in autism. Trends Cogn Sci. 2006;10(6):258–64. negativity and the P3a components of the auditory event-related potentials in 18. Fujita T, Yamasaki T, Kamio Y, Hirose S, Tobimatsu S. Parvocellular pathway autistic low-functioning subjects. Clin Neurophysiol. 2003;114(9):1671–80. impairment in autism spectrum disorder: evidence from visual evoked 44. Jansson-Verkasalo E, Ceponienè R, Kielinen M, Suominen K, Jäntti V, Linna S- potentials. Res Autism Spectr Disord. 2011;5(1):277–85. L, Moilanen I, Näätänen R. Deficient auditory processing in children with 19. Bertone A, Mottron L, Jelenic P, Faubert J. Enhanced and diminished Asperger syndrome, as indexed by event-related potentials. Neurosci Lett. visuo-spatial information processing in autism depends on stimulus 2003;338(3):197–200. complexity. Brain. 2005;128:2430–41. 45. Dunn MA, Gomes H, Gravel J. Mismatch negativity in children with autism 20. Baranek GT, David FJ, Poe MD, Stone WL, Watson LR. Sensory experiences and typical development. J Autism Dev Disord. 2008;38(1):52–71. questionnaire: discriminating sensory features in young children with 46. Isquith PK, Roth RM, Gioia G. Contribution of rating scales to the assessment autism, developmental delays, and typical development. J Child Psychol of executive functions. Appl Neuropsychol Child. 2013;2(2):125–32. Psychiatry. 2006;47(6):591–601. 47. Rosenthal M, Wallace GL, Lawson R, Wills MC, Dixon E, Yerys BE, 21. Leekam SR, Nieto C, Libby SJ, Wing L, Gould J. Describing the sensory Kenworthy L. Impairments in real-world executive function increase abnormalities of children and adults with autism. J Autism Dev Disord. 2007; from childhood to adolescence in autism spectrum disorders. 37(5):894–910. Neuropsychology. 2013;27(1):13–8. Høyland et al. Molecular Autism (2019) 10:10 Page 11 of 11

48. de Vries M, Geurts H. Influence of autism traits and executive functioning 74. Baron-CohenS,AshwinE,AshwinC,TavassoliT,ChakrabartiB.Talentin on quality of life in children with an autism spectrum disorder. J Autism autism: hyper-systemizing, hyper-attention to detail and sensory Dev Disord. 2015;45(9):2734–43. hypersensitivity. Philos Trans R Soc B Biol Sci. 2009;364(1522):1377–83. 49. Sbordone R. Neuropsychological tests are poor at assessing the frontal 75. Harms MB, Martin A, Wallace GL. Facial emotion recognition in autism lobes, executive functions, and neurobehavioral symptoms of traumatically spectrum disorders: a review of behavioral and neuroimaging studies. brain-injured patients. Psychol Injury Law. 2010;3(1):25–35. Neuropsychol Rev. 2010;20(3):290–322. 50. Sbordone RJ. The hazards of strict reliance on neuropsychological tests. 76. Silver CH. Sources of data about children’s executive functioning: review Appl Neuropsychol Adult. 2014;21(2):98–107. and commentary. Child Neuropsychol. 2014;20(1):1–13. 51. Gioia GA: BRIEF: behavior rating inventory of executive function: professional manual: psychological assessment resources; 2000. 52. Hovik KT, Egeland J, Isquith PK, Gioia G, Skogli EW, Andersen PN, Øie M. Distinct patterns of everyday executive function problems distinguish children with Tourette syndrome from children with ADHD or autism spectrum disorders. J Atten Disord. 2017;21(10):811–23. 53. Gioia GA, Isquith PK, Kenworthy L, Barton RM. Profiles of everyday executive function in acquired and developmental disorders. Child Neuropsychol. 2002;8(2):121–37. 54. Jonkman LM, Lansbergen M, Stauder JE. Developmental differences in behavioral and event-related brain responses associated with response preparation and inhibition in a go/nogo task. Psychophysiology. 2003;40(5):752–61. 55. Rutter M, Bailey A, Lord C: The social communication questionnaire: manual: Western Psychological Services; 2003. 56. Lord C, Rutter M, Le Couteur A. Autism diagnostic interview-revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. J Autism Dev Disord. 1994;24(5):659–85. 57. Berument SK, Rutter M, Lord C, Pickles A, Bailey A. Autism screening questionnaire: diagnostic validity. Br J Psychiatry. 1999;175:444–51. 58. Corsello C, Hus V, Pickles A, Risi S, Cook EH, Leventhal BL, Lord C. Between a ROC and a hard place: decision making and making decisions about using the SCQ. J Child Psychol Psychiatry. 2007;48(9):932–40. 59. Chandler S, Charman T, Baird G, Simonoff E, Loucas TOM, Meldrum D, Scott M, Pickles A. Validation of the social communication questionnaire in a population cohort of children with autism spectrum disorders. J Am Acad Child Adolesc Psychiatry. 2007;46(10):1324–32. 60. Wechsler D: Wechsler intelligence scale for children-WISC-IV: Psychological Corporation; 2003. 61. Roid GH, Miller LJ. Leiter international performance scale-revised (Leiter-R). Madrid: Psymtec; 2011. 62. Dumont R, Willis JO: Wechsler Abbreviated Scale of Intelligence. Encyclopedia of Special Education 1999. 63. Mueller A, Candrian G, Kropotov JD, Ponomarev VA, Baschera G-M. Classification of ADHD patients on the basis of independent ERP components using a machine learning system. Nonlinear Biomed Phys. 2010;4(Suppl 1):S1. 64. Benjamini Y, Hochberg YJJotrssSB: Controlling the false discovery rate: a practical and powerful approach to multiple testing. 1995:289–300. 65. Friedman NP, Miyake A. The relations among inhibition and interference control functions: a latent-variable analysis. J Exp Psychol Gen. 2004;133(1):101. 66. Luck SJ, Kappenman ES: ERP components and selective attention. The Oxford handbook of event-related potential components. New York: Published by Oxford University Press, Inc. 2012. p. 295–327. 67. Pellicano E, Burr D. When the world becomes ‘too real’: a Bayesian explanation of autistic perception. Trends Cogn Sci. 2012;16(10):504–10. 68. Ritter W, Simson R, Vaughan HG. Event-related potential correlates of two stages of information processing in physical and semantic discrimination tasks. Psychophysiology. 1983;20(2):168–79. 69. Johannes S, Munte TF, Heinze HJ, Mangun GR. Luminance and spatial attention effects on early visual processing. Brain Res Cogn Brain Res. 1995;2(3):189–205. 70. Jeste SS, Nelson CA 3rd. Event related potentials in the understanding of autism spectrum disorders: an analytical review. J Autism Dev Disord. 2009; 39(3):495–510. 71. Gjevik E, Sandstad B, Andreassen OA, Myhre AM, Sponheim E. Exploring the agreement between questionnaire information and DSM-IV diagnoses of comorbid psychopathology in children with autism spectrum disorders. Autism. 2015;19(4):433–42. 72. Lau-Zhu A, Fritz A, McLoughlin GJN, Reviews B. Overlaps and distinctions between attention deficit/hyperactivity disorder and autism spectrum disorder in young adulthood: Systematic review and guiding framework for EEG research. Neurosci Biobehav Rev. 2018. 73. McLoughlin G, Makeig S, Tsuang MT. In search of biomarkers in psychiatry: EEG-based measures of brain function. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics. 2014;165(2):111–21.